A HITL-Integrated Machine Learning Approach to Secure Drone Networks for IIoT Applications

The deployment of Unmanned Aerial Systems (UAS) in various industrial sectors like agriculture, military, and commercial operations is witnessing rapid growth, aligning with the broader trends in Industrial IoT (IIoT). With this escalation, ensuring security within UAS networks has emerged as a pivotal concern. Intrusion Detection Systems (IDS) play an indispensable role in maintaining the integrity and resilience of these systems. This paper introduces a novel framework for intrusion detection tailored to the specialized needs of IIoT, aimed at fortifying the security of UAS networks within industrial applications. By integrating Human-In-The-Loop (HITL) strategies with machine learning algorithms, the framework enhances detection performance, capitalizing on the insights and expertise of human analysts. This involvement allows them to actively participate in training the machine learning model, thereby amplifying its efficiency. To gauge the effectiveness of our HITL-ML approach, we carried out a comparative analysis against traditional machine learning-based intrusion detection systems using metrics pertinent to industrial contexts. The experimental results compellingly show that our HITL-ML framework outperforms conventional methods, affirming that the fusion of human expertise with machine learning models offers a more robust and streamlined solution for identifying and counteracting threats in industrial UAS networks.